Removing input type from schema does not remove generated type

[dwbutler]

My Netflix DGS project is using com.netflix.graphql.dgs.codegen:graphql-dgs-codegen-shared-core=5.4.0

On my local machine, I was developing a new schema with a new input type. Netflix DGS generated the input type, I wrote code against it, everything was good.

Then, I decided to rename the input type to something else. I regenerated the schema using the generateJava task, everything was good. But I forgot to go into my application code and use the new input type name. Everything compiled so I thought it was good, so I pushed up to CI. It failed on the CI server because the old type name didn't exist anymore.

I would expect Netflix DGS to regenerate a clean set of classes each time by wiping out ALL of the existing generated code and generating all new files every time. Or at the very least, detecting code files that don't belong anymore and deleting them.

[srinivasankavitha]

Can you please try with the latest version 6.0.2?Sent from my iPhoneOn Aug 24, 2023, at 1:54 PM, David Butler ***@***.***> wrote: My Netflix DGS project is using com.netflix.graphql.dgs.codegen:graphql-dgs-codegen-shared-core=5.4.0 On my local machine, I was developing a new schema with a new input type. Netflix DGS generated the input type, I wrote code against it, everything was good. Then, I decided to rename the input type to something else. I regenerated the schema using the generateJava task, everything was good. But I forgot to go into my application code and use the new input type name. Everything compiled so I thought it was good, so I pushed up to CI. It failed on the CI server because the old type name didn't exist anymore. I would expect Netflix DGS to regenerate a clean set of classes each time by wiping out ALL of the existing generated code and generating all new files every time. Or at the very least, detecting code files that don't belong anymore and deleting them. —Reply to this email directly, view it on GitHub, or unsubscribe.You are receiving this because you are subscribed to this thread.Message ID: ***@***.***>

[valterfab]

Hello, I can confirm that this is still happening on v6.0.2